J. Phycol. *, ***–*** (2018) © 2018 Phycological Society of America DOI: 10.1111/jpy.12766
BAFFINELLA FRIGIDUS GEN. ET SP. NOV. (BAFFINELLACEAE FAM. NOV., CRYPTOPHYCEAE) FROM BAFFIN BAY: MORPHOLOGY, PIGMENT PROFILE, PHYLOGENY, AND GROWTH RATE RESPONSE TO THREE ABIOTIC FACTORS1
Niels Daugbjerg,2 Andreas Norlin Marine Biological Section, Department of Biology, University of Copenhagen, Universitetsparken 4, Copenhagen Ø DK-2100, Denmark and Connie Lovejoy Departement de Biologie, Universite Laval, 1045 avenue de la Medecine, Quebec, Quebec, G1V 0A6, Canada
Twenty years ago an Arctic cryptophyte was Abbreviations: BA, Bayesian analysis; BS, bootstrap isolated from Baffin Bay and given strain number support; Cr-PC, cryptophyte-phycocyanin; Cr-PE, CCMP2045. Here, it was described using cryptophyte-phycoerythrin; ML, maximum likeli- morphology, water- and non-water soluble pigments hood; PP, posterior probability and nuclear-encoded SSU rDNA. The influence of temperature, salinity, and light intensity on growth rates was also examined. Microscopy revealed = typical cryptophyte features but the chloroplast Cryptophytes ( cryptomonads) are ubiquitous in color was either green or red depending on the marine and freshwater ecosystems worldwide and a few species have been recorded to form blooms light intensity provided. Phycoerythrin (Cr-PE 566) was only produced when cells were grown under (e.g., Laza-Martınez 2012, Supraha et al. 2014, and low-light conditions (5 lmol photons m 2 s 1). references therein). However, they also reside in Non-water-soluble pigments included chlorophyll a, more extreme environments, for example, soil (Paulsen et al. 1992), snow (Javornick y and Hind ak c2 and five major carotenoids. Cells measured 8.2 3 5.1 lm and a tail-like appendage gave them a 1970), and inside ikaite columns (Ikka fjord, South- comma-shape. The nucleus was located posteriorly west Greenland; Kristiansen and Kristiansen 1999, and a horseshoe-shaped chloroplast contained a N. Daugbjerg, unpub. data). The currently accepted single pyrenoid. Ejectosomes of two sizes and a taxonomy of Cryptomonada includes two classes: nucleomorph anterior to the pyrenoid were the photoautotrophic Cryptophyceae (note: Chilomo- discerned in TEM. SEM revealed a slightly elevated nas is heterotrophic but this state is most likely sec- vestibular plate in the vestibulum. The inner ondarily derived) and the phagotrophic periplast component consisted of slightly Goniomonadea. Until recently, the latter class con- overlapping hexagonal plates arranged in 16–20 sisted of only a single genus Goniomonas and some oblique rows. Antapical plates were smaller and environmental sequences, but Shiratori and Ishida their shape less profound. Temperature and salinity (2016) have described the second heterotrophic studies revealed CCMP2045 as stenothermal and cryptomonad (Hemiarma marina). euryhaline and growth was saturated between 5 and The species diversity of Cryptophyceae has been 20 lmol photons m 2 s 1. The phylogeny based studied for more than 180 years as Ehrenberg dis- on SSU rDNA showed that CCMP2045 formed a covered the first cryptophyte (Cryptomonas ovata)in distinct clade with CCMP2293 and Falcomonas sp. 1831. However, the first detailed descriptions and isolated from Spain. Combining pheno- and drawings were published 7 years later (Ehrenberg genotypic data, the Arctic cryptophyte could not be 1838). According to AlgaeBase, as of May 2018, placed in an existing family and genus and there are 41 genera listed in the Cryptophyceae (Guiry and Guiry 2018) and of these 21 are mono- therefore Baffinellaceae fam. nov. and Baffinella = frigidus gen. et sp. nov. were proposed. typic ( 53.7%). At the biochemical level, cryptophytes possess a Key index words: Arctic flagellates; cryptophytes; unique combination of pigments that include growth rates; phylogeny; taxonomy; ultrastructure chlorophylls a and c2, one of two phycobiliproteins and the cryptophyte specific carotenoid alloxanthin, alongside other common carotenoids (Hill and Rowan 1989, Cerino and Zingone 2007, Egeland 1 Received 17 April 2018. Accepted 27 June 2018. First Published 2016). The two phycobiliproteins found in crypto- Online 25 July 2018. 2Author for correspondence: e-mail [email protected]. phytes are phycoerythrin and phycocyanin, and Editorial Responsibility: L. Graham (Associate Editor) these have different structural configurations
1 2 NIELS DAUGBJERG ET AL. identified by their absorption peaks. Some of these with existing families. Thus, a new family Baffinel- configurations are unique (Hill and Rowan 1989, laceae fam. nov. was also described. Novarino 2012). For each genus, only one type of phycobiliprotein is present except for Hemiselmis that has been shown to contain species with either MATERIALS AND METHODS Cr-PC 615 and Cr-PE 555 (Hill and Rowan 1989). Culture. Strain CCMP2045 was isolated from a water sam- 0 As some families possess the same phycobiliprotein ple collected in Baffin Bay (June 23, 1998; 76°19 15″ N, ° 0 ″ (Clay et al. 1999), it has been discussed if the pho- 75 49 02 W; Fig. 1) at a depth of 11 m. A sub-culture of this tosynthetic pigment profile can be used to distin- strain was sent to the Marine Biological Section at University of Copenhagen where it has been kept at 4°C in L1 medium guish taxa at the family level (Fritsch 1935, (Guillard and Hargraves 1993) with a salinity of 30 and a Pringsheim 1944, Butcher 1967, Hill and Rowan light:dark cycle of 16:8 h. 1989). Microscopy. Live cells were studied using a Carl Zeiss Axio At the light microscopical level, the gross mor- Imager.M2 with a 639 oil immersion lens. Micrographs were phology of cryptophytes is somewhat similar: droplet taken with a Zeiss AxioCam HRc digital camera (Zeiss, Ober- cell shape and two flagella that emerge from the kochen, Germany). Chlorophyll fluorescence was recorded subapical end. Thus, most taxonomic differences using a Zeiss AxioCam MRc digital camera and the filter set 09 (excitation BP450-490, emission LP515). Length and width relate to the ultrastructure of the periplast compo- of 59 cells were measured using Zen image acquisition soft- nent, position of the nucleomorph, appearance of ware from Zeiss. Means and standard deviations were calcu- the rhizostyle, and the furrow-gullet complex (Clay lated using IBM SPSS statistics (ver. 24, Armonk, NY, USA). et al. 1999). For transmission electron microscopy, 5 mL of a dense cul- In the Arctic, cryptomonads have been shown to ture was fixed in 5 mL of 4% glutaraldehyde in 0.2 M cacody- contribute to both pelagic and sympagic communi- late buffer containing 0.5 M sucrose for 1 h and 25 min. The ties of microalgae. They are also part of the spring fixed cells were pelleted by centrifugation and the super- natant except for 2 mL was removed. Two milliliter 0.2 M bloom when the sea ice melts and nutrients are cacodylate buffer containing 0.5 M sucrose was added. After released to the open water column (Mikkelsen 20 min, 2 mL of the supernatant was removed and replaced et al. 2008, Vallieres et al. 2008, Poulin et al. 2011, with 2 mL 0.2 M cacodylate buffer without sucrose for Coupel et al. 2012). However, very few studies have 20 min. The mix was centrifuged at 1,201 g for 10 min after addressed the taxonomy of Arctic cryptophytes. each of these steps. For post-fixation 2 mL of the supernatant Recently, the species diversity of Arctic marine pro- tists including cryptomonads was studied by molec- ular techniques (e.g., clone libraries and high throughput sequencing; Sørensen et al. 2012, Mar- quardt et al. 2016, Stecher et al. 2016). These stud- ies have provided detailed information on operational taxonomic units, but for many of the determined ribotype sequences we have little understanding of the corresponding morphos- pecies. In this study, a cryptophyte was isolated from material collected in Baffin Bay in Arctic Canada (1998), and later deposited at Provasoli-Guillard National Center for Marine Micro Algae and Micro- biota in Maine, USA (Potvin and Lovejoy 2009), where it was given the strain number CCMP2045. The aim of this study was to describe this strain using a polyphasic approach, which included a mor- phological characterization using LM, TEM and SEM. The composition of both water and non-water- soluble pigments was analyzed to describe the pro- file of photosynthetic marker pigments. To infer the phylogeny of CCMP2045, the nuclear-encoded SSU rDNA sequence was compared to 50 other cryp- tomonads. Finally, growth rates were estimated through autecological experiments elucidating salin- ity and temperature tolerance limits as well as growth rates at different light intensities. Based on these results, a new genus and species Baffinella frigi- dus gen. et sp. nov. was proposed to accommodate for strain CCMP2045. The combination of morpho- FIG. 1. Map of sampling location (*)ofBaffinella frigidus gen. logical features and phycobiliproteins did not fit et sp. nov. in Baffin Bay. BAFFINELLA FRIGIDUS, AN ARTIC MARINE CRYPTOPHYTE 3
was removed and 2 mL, consisting of 0.5 mL 4% OsO4 and previously been determined by Potvin and Lovejoy. The 1.5 mL 0.2 M cacodylate buffer was added for 1 h before sequences were retrieved from GenBank (accession number being rinsed with 0.2 M cacodylate buffer. For dehydration, a GQ375264 and GQ375265, respectively) and added to an series of ethanol was used, 20 min for each concentration alignment consisting of 50 other SSU rDNA sequences of (30%, 50%, 70%, 90%, and 100% in molecular sieves) with cryptomonads including 18 genera and 38 taxa identified to 100% in molecular sieves used twice. The pellet was rinsed species. The data matrix was edited using the JALVIEW twice in propylene oxide before embedding and left over- sequence editor (ver. 2.9.0b2; Waterhouse et al. 2009) follow- night in a mix of 1:1 propylene oxide and Spurr’s resin. The ing an alignment based on ClustalW as implemented in the next day Spurr’s resin was replaced by fresh Spurr’s resin for software. The data matrix consisted of 1942 base pairs includ- 6 h before the pellet was embedded in a cast with fresh ing introduced gaps. Phylogenetic reconstructions were based Spurr’s resin and polymerized at 70°C overnight. Thin sec- on BA and ML analyses. BA was performed using MrBayes tions were cut with a diamond knife and stained for transmis- (ver. 3.2.5 X64; Ronquist and Huelsenbeck 2003) on a desk- sion electron microscopy using uranyl acetate and lead top computer where 5 9 106 generations were run and a tree citrate. The microscope used was a JEM 1010 electron micro- was sampled every 1,000 generations. The burn-in value was scope (JEOL Ltd., Tokyo, Japan) and micrographs were taken evaluated using the lnL value plotted as a function of genera- using a GATAN Orius SC1000 digital camera (Gatan, Pleasan- tions. The lnL values converged after 101,000 generations ton, CA, USA). The accelerating voltage was 80 kV. and 101 trees were discarded, leaving 4,900 trees for generat- For scanning electron microscopy, 3.5 mL of the culture ing a 50% majority-rule consensus tree in PAUP* (Swofford was fixed in 3.5 mL 4% glutaraldehyde in 0.2 M cacodylate 2002). ML analysis was run using the PhyML software (Guin- buffer containing 0.5 M sucrose for 1.5 h, adding 1 mL of don et al. 2010) on a desktop computer with the parameter 4% OsO4 in seawater after 10 min. The cells were placed on settings obtained from jModelTest (ver. 2.1.3; Darriba et al. a5lm isopore filter and rinsed in 0.2 M cacodylate buffer 2012). GTR+I+G was chosen as the best fit model for the data containing 0.5 M sucrose followed by 0.2 M cacodylate buffer matrix (gamma = 0.568 and p-invar = 0.464). BS were calcu- containing 0.25 M sucrose and finally 0.2 M cacodylate buffer lated from 500 replications and mapped on the tree from BA without sucrose, for 10 min each. Dehydration was done with alongside the PP values. an ethanol series (30%, 50%, 70%, 90%, and 99.9%) for The katablepharid Roombia truncata was used to root the 10 min each. Dehydration was finished with 100% ethanol in tree as Burki et al. (2012) in a large-scale phylogenetic study molecular sieves for 30 min twice before the sample was criti- had shown this taxon to form a sister to the cryptomonads. cal-point-dried. The isopore filter was attached to an alu- Autecological experiments. For salinity and temperature, the minum stub and sputter coated with gold/palladium in a autecological experiments were performed in 60 mL Nunc JEOL JFC-2300HR sputter coater before being observed in a flasks. Triplicate flasks were placed in 24 L glass tanks con- JEOL JSM-6335F field emission scanning electron microscope taining demineralized water, circulated by an aquarium with a secondary electron detector and running at 12 kV pump. A cooling system controlled by a custom made relay (JEOL Ltd.). switched cooled water on/off and thus ensured accurate con- Pigment analyses. Extraction of the water-soluble pigment trol of pre-set experimental temperatures ( 0.1°C). For light (phycobiliprotein) was based on Lawrenz et al. (2011) and experiments, the Nunc flasks were placed in a 4°C walk-in cli- briefly described here. An exponentially growing live sample mate room, at areas where the desired light intensities were (5–10 mL) was centrifuged for 5 min at 3075 g and the marked. Light measurements were made using a spherical supernatant was removed. A volume of 2.5 mL of PBS buffer PAR sensor (Walz, ULM-500; Heinz Walz GmbH, Effeltrich, (pH 7.2 or 7.5) was added and cells were disrupted by two Germany). The light experiments were conducted at a light:- freeze-thaw cycles (i.e., 10 min at 80°C followed by thawing dark cycle of 16:8 h, because a shared climate room was used. at room temperature). The extraction lasted at least 20 h at pH was measured regularly (often every second day) to moni- 4°C. Finally, a subsample of 1.2 mL was put in a 1.5 mL tor increases due to photosynthetic activity. All experiments Eppendorf tube and centrifuged for 15 min at 12,202g. The used a start concentration of ~1,000 cells mL 1. Samples for supernatant was analyzed by scanning the spectrum between cell abundance measurements were collected shortly after the 350 and 800 nm in a Shimadzu UV180 spectrophotometer start of the experiments (time = t ). Due to a linear correla- 0 (Holm & Halby A/S, Brøndby, Denmark). Extraction of non- tion between fluorescence of chlorophyll a and cells mL 1 water-soluble pigments for HPLC were extracted in 95% when grown at the same light intensity, we converted fluo- methanol, and separated with a Accela 600 HPLC system rometer readings to cell abundances based on a standard (Thermo Scientific, San Jose, CA, USA) on a reverse-phase curve. See Figure S1 in the Supporting Information for the Hypersil Gold C-8 column and a solvent gradient containing standard curve used in the temperature and salinity experi- methanol, aqueous pyridine, acetone, and acetonitrile (Zap- ments and Figure S2 in the Supporting Information for the ata et al. 2000). An Accela PDA detector recorded chro- standard curves used in the light experiments. Samples were matograms at 450 nm and pigment spectra over the taken every second day by pipetting 1,140 lL from each repli- wavelengths 350–800 nm. A Finnigan Surveyor FL Plus fluo- cate into 1.5 mL Eppendorf tubes. Raw fluorescence units rescence detector with excitation of 440 nm and emission at were measured using a Trilogy fluorometer equipped with 650 nm (optimized for Chl a) was also used to identify and the blue module (Trilogy, Turner Designs Instruments, Sun- quantitate Chl pigments. The system was calibrated by nyvale, CA, USA). The first replicate (A series) of each treat- repeated injections of 30 pigment standards (Sigma Aldrich ment was used to calculate standard curves by fixing it with (St. Louis, MO, USA) or DHI (Hørsholm, Denmark) LAB Lugol’s iodine (5% final concentration) and counting cell products). ChromeQuest software was used to identify and numbers using 1.0 mL Sedgewick Rafter chambers (Pyser-SGI quantitate the concentration of the pigments. The photodi- Limited, Edenbridge, UK). ode array spectrum of each peak was checked against the ref- Temperature experiments. The growth rates were examined erence spectra of the standard or against the reference at 4.0°C, 6.0°C, 8.0°C and 10.0°C and at a salinity of 35. spectra in Roy et al. (2011) for pigments for which there are The light intensity was 100 lmol photons m 2 s 1 and no standards. the light:dark cycle 24:0 h to simulate the Arctic summer Phylogenetic inference. The phylogenetic inference of conditions. CCMP2045 was based on the nuclear-encoded SSU rDNA Salinity experiments. These were conducted at salinities of gene, alongside the sequence of CCMP2293, which had 0, 5, 10, 15, 20, 25, and 30 and all at 4.0 0.1°C. An 4 NIELS DAUGBJERG ET AL. acclimation period of 3 d was included at salinities of 10 used side; flagella inserted at one-third cell length for experiments at 0, 5, and 10, respectively; 20 used for from the apex. experiments at 15 and 20, respectively; and 30 used for exper- Etymology: Frigidus from latin meaning cold, refer- iments at 25 and 30, respectively. The light setting was identi- cal to that used in the temperature experiments. The same ring to its Arctic origin. set of flasks and therefore the same data were used for the Holotype: An epon-embedded sample of strain temperature experiment at 4°C and salinity experiment at 35. CCMP2045 has been deposited at the National His- Light experiments. These were conducted at 5, 20, 40, 60, tory Museum, University of Copenhagen (C), acces- 80, and 100 lmol photons m 2 s 1 at 4°C and a salinity of sion no. C A92087. 30. The light:dark cycle was 16:8 h. GenBank accession number: GQ375264, represents Calibration of growth rates. Growth rates were calculated for the nuclear-encoded SSU rDNA sequence. each replicate and the mean was based on three replicates Authentic culture: CCMP2045 at the Provasoli-Guil- (n = 3). The maximum growth rate (l) was defined as the number of cell divisions d 1 and calculated as: lard National Center for Marine Algae and Micro- biota. Type locality: Baffin Bay (76°19015″ N, 75°49002″ ln Nt2 l ¼ Nt1 W), Arctic Canada. t2 t1 Morphology, color, and swimming behavior in LM. Cells measured 5.6–11 lm in length (average: where Nt1 and Nt2 were the cell abundances at the start (t1) l = – l and end (t ), respectively. 8.2 1.2 m, n 59) and 4.3 6.1 m in width (av- 2 l = – Statistics. Comparisons of mean growth rates by one-way erage: 5.1 0.4 m, n 59; Fig. 2, A J). Most cells ANOVA tests were performed using Graphpad Prism (ver. 7) had a comma-shaped appearance due to a small and GraphPad Prism (ver. 6, GraphPad Software, La Jolla, tail-like appendage in the antapical end, others were CA, USA) was used to plot the data and do regression analy- bluntly rounded (Fig. 2, A, B and D). In optical ses. cross-section, cells were spherical (Fig. 2J). Two unequally long flagella were inserted in an apical RESULTS depression (the gullet, Fig. 2, A-C, F, H and K). Rows of large ejectosomes were observed close to We propose the following taxonomy for strain the gullet (Fig. 2, A, B and E). A large vacuole was CCMP2045. present in the apical end (Fig. 2H), whereas the Class Cryptophyceae. Order Cryptomonadales. nucleus was located in the antapical end (Fig. 2, C Chloroplasts with phycobiliprotein Cr-phycoerythrin and G). The structure of the periplast component 566 (Cr-PE III), leucoplasts present in some. surrounding the cell could be discerned in the light Description: Baffinellaceae Daugbjerg & Norlin microscope (Fig. 2F). The single chloroplast was fam. nov. seen to extend in almost the full length of the cell Furrow-gullet complex present, without stoma; (Fig. 2M). In cross-section, the two chloroplast lobes nucleomorph positioned anterior to the pyrenoid; formed a horseshoe-shape (Fig. 2, J and L). A single inner periplast component with partly overlapping pyrenoid was observed in connection with the hexagonal plates; subapical vestibular plate present; chloroplast and it was surrounded by a starch sheet rhizostyle present. (Fig. 2I). The pyrenoid was located in the middle of Description: Baffinella Norlin & Daugbjerg gen. the cell (Fig. 2G). A refractive crystal-like structure nov. was present in many cells (Fig. 2, A–C). The ejecto- Two unequal flagella protruding from a furrow-gul- somes discharged when cells experienced elevated let system in the anterior end; vestibular plate located heating from prolonged observations in the light on dorsal side of gullet; a single horseshoe-shaped microscope (Fig. 2N). Live cells swam while rotating chloroplast with a single pyrenoid surrounded by a around their longitudinal axis. Depending on the starch sheet; apical end rounded and antapical end light intensity provided the color of CCMP2045 was acute or bluntly rounded; hexagonal periplast plates either green or red. Live cells grown at low light less profound in shape at the antapical end. (5 lmol photons m 2 s 1) were red (Fig. S3, A– Type species: B. frigidus sp. nov. C in the Supporting Information), whereas cells Etymology: from the name of the collection area, were green if grown at light intensities ≥10 lmol the Baffin Bay, with the Latin, feminine suffix ella, photons m 2 s 1 (Fig. S3, D–F). Interestingly, meaning small. CCMP2045 has been identified as Rhodomonas sp. in Description: Baffinella frigidus Norlin and Daugb- the culture collection at Bigelow Laboratory for jerg sp. nov. Ocean Sciences (http://ncma.bigelow.org/cc Comma-shaped in lateral view, spherical in mp2045); probably due to low light conditions pro- cross-section; dorsal side convex, ventral side vided. more straight; live cells 5.6–11 lmlongand4.3– l External morphology in SEM. The apical vestibulum 6.1 m wide; gullet deep, lined with rows of large was part of a furrow-gullet complex (Fig. 3, A, B ejectosomes; pyrenoid bisected by cytoplasmic and D). The two flagella emerged on the right side tongue; mid-dorsal band present at antapical end; of the gullet (Fig. 3, A and B). The vestibular plate nucleus in antapical end, slightly to the dorsal formed a tongue-like structure located in the upper BAFFINELLA FRIGIDUS, AN ARTIC MARINE CRYPTOPHYTE 5